ABSTRACT The proposed research program will focus on two pharmacogenetic challenges that hold great potential for improving patient therapy. 1) The first challenge is to define the clinical importance of known pharmacogenetic associations, as related to drug toxicity and therapeutic failure?common outcomes of any therapy. The genetic contribution to variability in pharmacokinetics and pharmacodynamics, and thus potentially to differences in drug response, is well recognized. However, despite intensive research, little of this work has translated to clinical practice. A critical barrier is that the effect of genotype on meaningful patient outcomes in clinical practice is not known. Ideally, large randomized controlled trials would define the effect of genotype on clinical outcomes, and a few such studies are underway. However, given the expense of such trials, the many genotypes, drugs, and outcomes of interest, as well as the difficulties extrapolating from precise clinical trials to imprecise clinical practice, this approach is limited. Consequently, the path forward to translate science into practice is unclear. Accordingly, we propose a novel, cost effective approach: to use a de-identified electronic medical record (EHR) linked to a DNA biobank with >200,000 patients (BioVU) to define the clinical importance of variation in genes affecting drug metabolism or response. The long-term goal of this area of work is to develop and implement methods to define the importance of genetic variation on the outcomes of drug therapy in real world clinical practice. 2) The second challenge is to use genetics to predict unexpected toxicities and benefits of drugs. Over time, most drugs newly introduced to the market are found to have additional therapeutic indications and also unexpected toxicities. A critical barrier is that traditional post-marketing approaches to define these effects often require decades of study. During this time patients would accrue unwanted currently unknown adverse effects and forgo potential off-target benefits. Genetic approaches can provide information to speed this process. The mechanism of action of some drugs (e.g., ezetimibe that inhibits Nieman-Pick C1-like 1 (NPC1L1)) are mimicked by variations in genes (NPC1L1). By studying individuals who carry these variants and determining their outcomes in large EHR databases using a technique called phenome-wide association studies (PheWAS) followed by fine-phenotyping we can infer potential outcomes when patients are exposed to the drug. The long-term goal of this area of work is to develop and implement methods to use genetic information to discover unexpected benefits and risks of drugs. The two pharmacogenetic challenge areas that will be the foundation of the research program have high public health impact, not only in translating basic science to improved patient care for the drugs studied, but also in providing new approaches for testing the clinical importance of a range of drug/genotype questions.